On Mon, 2011-11-21 at 11:21 +0100, Eric Dumazet wrote:

Le lundi 21 novembre 2011 à 11:15 +0100, Eric Dumazet a écrit :

quoted

Hmm, trinity tries to crash decnet ;)

Maybe we should remove this decnet stuff for good instead of tracking
all bugs just for the record. Is there anybody still using decnet ?

For example dn_start_slow_timer() starts a timer without holding a
reference on struct sock, this is highly suspect.

[PATCH] decnet: proper socket refcounting

Better use sk_reset_timer() / sk_stop_timer() helpers to make sure we
dont access already freed/reused memory later.

Reported-by: Sasha Levin <redacted>
Signed-off-by: Eric Dumazet <redacted>

Hmm, I forgot to remove the sock_hold(sk) call from dn_slow_timer(),
here is V2 :

[PATCH] decnet: proper socket refcounting

Better use sk_reset_timer() / sk_stop_timer() helpers to make sure we
dont access already freed/reused memory later.

Reported-by: Sasha Levin <redacted>
Signed-off-by: Eric Dumazet <redacted>
---

[snip]

Applied locally and running same tests as before, will update with
results.

-- 

Sasha.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>